An efficient lagrangian trajectory model for predicting complex turbulent two-phase flows using the curvilinear coordinates

Abstract

In this article, a numerical procedure is developed to extend the SPEED (Stochastic-Probabilistic Efficiency Enhanced Dispersion) model to the non-orthogonal curvilinear coordinates for the efficient prediction of dispersed turbulent two-phase flows with complex geometries. The partial-transformation approach using the Cartesian velocity vectors is employed to compute the fluid and dispersed flow fields. The Lagrangian trajectory model accounts for both the stochastic and probabilistic effects of discrete particle dispersion induced by fluid turbulence. A robust numerical algorithm is developed to determine the probability distribution in irregular Eulerian control volumes. The extended SPEED model is validated against a 90 deg turbulent bend flow laden with solid particles. It is found that the computational efficiency is greatly enhanced by using a very fewer number of particle trajectories than the conventional stochastic model while achieving the noise-free computational results.